The surest sign that a technology has arrived is not the keynote; it is the expense report. Meta’s Muse is a hit, Instinct remains buzzy, and the era of the personal AI agent — software that connects to your bank, email and calendar and acts on the internet for you — appears to be upon the executive class. Business Insider asked bosses at Anduril, KPMG, Adobe, Uniphore and Zendesk how they use the tools at home. The answers ran from drudgery to dog care, with one common caveat: a human keeps the final say before anything is sent or bought.
Jeff Miller, Anduril’s chief marketing officer, calls personal AI “wild.” “I’m on Muse, ChatGPT, you name it,” he said. The executive at a company that builds autonomous defence technology recently used AI to clear six months’ worth of expense reports, hire a new dog walker and sharpen his fitness routine. “I’m using it literally right now to negotiate new insurance rates for me,” he told Business Insider. He grants the tools access to his personal bank details and email, but not the right to act unsupervised: “I have faith in the security from a personal standpoint, but ultimately, when I send an email, when I buy something, I’m the one hitting yes or no.”
Eric Snowden, senior vice-president of design at Adobe, used Muse to book a restaurant in Amsterdam during the HumanX conference and, domestically charming, a flu shot for his return. What impressed him more was initiative: “It was smart enough to realize I was doing a renovation on my apartment, flagged an email from my contractor that I didn’t see, and was like, ‘you need to sign these important documents,’” he said. Mr Snowden, who is helping oversee Adobe’s transition as AI reshapes creative work, said he might even pay a subscription for something like Muse. “It’s the first tool I’ve used where I could see a non-technical person really using this.”
Umesh Sachdev, chief executive of Uniphore, builds agents for a living and has wired Muse to his social-media accounts to warn him when topics or figures stir the discourse. “I don’t have FOMO of what I’m missing,” he said. “What has happened in the world in the last three hours I’ve been in Amsterdam? What has Trump done? Muse is doing that for me with X, LinkedIn, and so forth.”
Todd Lohr, KPMG’s AI chief, plays with the tools daily despite not being an engineer by background; trip-planning is the most common household use, and he has been “tinkering around” with automating security settings and cameras. Mr Lohr, named in September to head KPMG’s new Silicon Valley-inspired AI division, reports no mishaps so far. “I like to think I know how to prompt to not hallucinate,” he said — but he supervises home agents as he does work ones: “You need to build agents to check your agents. You need to check the outputs.”
Tom Eggemeier, Zendesk’s chief executive, is turning his company into one where AI agents do ever more of the work, and has tested Muse at home with immediate effect: the agent haggled better deals on his phone plan and his internet, saving him about $2,000 a year. With a bit more tuning, he reckons, it will save more still.
Taken together, the confessions describe a technology past the demo stage and into genuine usefulness — booking, chasing, haggling and summarising well enough to reclaim hours. Yet the pattern beneath enthusiasm is caution. The same week produced a reminder of why: one early Muse user found his agent had handed out his home address and accepted a cut-price offer without asking. The executives’ verdict is less a rave review than a job description. The agent is a diligent, tireless and occasionally dim assistant; the humans, for now, are keeping the signing pen.